CloudConsole

CloudWatch usage types decoded

Updated 17 November 2026 · How we write and check guides

Short answer

DataProcessing-Bytes is the CloudWatch usage type for log data ingested into Standard-class log groups, and it is usually the biggest CloudWatch line. Other common ones are TimedStorage-ByteHrs (log storage), DataScanned-Bytes (Logs Insights queries), MetricMonitorUsage (custom and detailed-monitoring metrics), GMD-Metrics (GetMetricData calls) and AlarmMonitorUsage (alarms). Each carries a Region prefix, such as USE1- or EU-.

The CloudWatch usage types you will see most: EUW2-DataProcessing-Bytes = Region, log ingestion
Where this charge comes from, using the dimensions and usage types named on this page.

How the names are built

AWS documents the format as <Region>-CW:<Feature> or <Region>-<Feature>. So EU-CW:GMD-Metrics is GetMetricData requests in eu-west-1, and USE1-DataProcessing-Bytes is log ingestion in us-east-1. Some free tier usage is tracked under a Global- prefix. Note that the Logs lines reuse names other services also use (TimedStorage-ByteHrs also exists in S3), so filter by Service = CloudWatch first.

The usage types you are most likely to see

Usage typeFeatureWhat drives it
DataProcessing-BytesLogs ingestion, Standard classGB written by applications
DataProcessingIA-BytesLogs ingestion, Essentials (Infrequent Access) classGB written
VendedLog-BytesVended logs into CloudWatch LogsVPC flow logs, Route 53 resolver logs and similar
S3-Egress-Bytes, FH-Egress-BytesVended logs delivered to S3 or FirehoseGB delivered
TimedStorage-ByteHrsLog storageRetention x volume
DataScanned-BytesLogs InsightsGB scanned by queries, including dashboards that run them
MetricMonitorUsageCustom metrics, detailed monitoring, EMF and metric-filter metricsNumber of metrics, prorated hourly
RequestsAPI calls such as PutMetricData, GetMetricStatistics, ListMetricsCall volume above the free tier
GMD-MetricsGetMetricDataMetrics requested, charged from the first call
MetricStreamUsageMetric streamsMetric updates streamed
AlarmMonitorUsage, HighResAlarmMonitorUsageMetric alarmsAlarm metrics, prorated hourly
DashboardsUsageHourDashboardsDashboards over the free allowance

Less common lines that still surprise people

  • Logs-LiveTail: Live Tail session minutes above the monthly free allowance; a tab left open on a busy log group keeps the session running.
  • DataProtection-Bytes: log data protection policies scanning and masking sensitive data, billed per GB scanned.
  • ContributorInsightRules and ContributorInsightEvents: Contributor Insights rules and the log events they match.
  • Canary-runs: Synthetics canaries, billed per run, so a one-minute schedule adds up across many canaries.
  • DatabaseInsights-vCPU-Hours or -ACU-Hours: Database Insights in Advanced mode.
  • ObservationUsage and MetricsUsage: Container Insights with enhanced observability for EKS and ECS.

Break CloudWatch spend down

aws ce get-cost-and-usage --region us-east-1 \
  --time-period Start=2026-11-01,End=2026-11-17 \
  --granularity MONTHLY --metrics UnblendedCost \
  --filter '{"Dimensions":{"Key":"SERVICE","Values":["AmazonCloudWatch"]}}' \
  --group-by Type=DIMENSION,Key=USAGE_TYPE Type=DIMENSION,Key=OPERATION \
  --query 'ResultsByTime[0].Groups[].[Keys[0],Keys[1],Metrics.UnblendedCost.Amount]' \
  --output text | sort -k3 -gr | head -20

The Operation adds detail: PutLogEvents for ingestion, HourlyStorageMetering for storage, StartQuery for Logs Insights, and for metrics MetricStorage:AWS/Logs-EMF (embedded metric format) or MetricStorage:AWS/CloudWatchLogs (metric filters). If the service filter returns nothing, check the exact service name with aws ce get-dimension-values --dimension SERVICE.

Where to go next for each line

A common misreading is blaming metrics for a CloudWatch bill that is really log ingestion. Check DataProcessing-Bytes first; it dominates for most accounts.

Questions

What is DataProcessing-Bytes in CloudWatch?

The usage type for log data ingested into Standard-class CloudWatch Logs log groups, billed per GB.

What is CW:GMD-Metrics?

Charges for GetMetricData API requests, billed by the number of metrics requested; these calls are not covered by the free API request allowance.

What is MetricMonitorUsage?

Custom metrics, detailed monitoring metrics, embedded metric format metrics and metric-filter metrics, each billed per metric per month, prorated by the hour.

When this happens away from your desk

CloudWatch has more usage types than most services, and the names are cryptic. CloudConsole, a phone app for AWS, shows spend by usage type and explains each CloudWatch line in plain words, so you can tell logs from metrics from your phone.

About the app → · Download on the App Store

Sources